A few days ago I argued QE may re-appear in both Europe and the United States, possibly within a year. A resumption of balance sheet expansion by the Federal Reserve emerges from the mechanics of financial plumbing, which now require precautionary liquidity buffers from the commercial banking sector (macro-demand). The earlier piece did not address the wider demand for reserves stemming from the settlement demand. This article looks at settlement demand. Three key takeaways emerge.
There is long-running evidence that turnover of reserves via FedWire settlement system has an impact on nominal GDP – and almost certainly inflation.
The macro-demand for reserve balances remains relatively sanguine (see previous article). The settlement demand for reserve balances is much less clear, but evidently has an immediate relevance to turnover and settlement in securities markets, repo and, thus, to the wider economy.
Ultimately, real-time settlement of securities, and especially repo, will likely determine the timing of the Fed’s need to add to its balance sheet.
Our earlier piece looked at liquidity demand as a function either of a percentage of GDP, or percentage of Total Assets of commercial banks. A more subtle assessment of liquidity demand would account for the transmission of liquidity to the total economy. One way to gauge this is the interaction of reserves, FedWire settlement system and the wider investment universe. FedWire settles funds and securities instantaneously across the commercial banking system, plus assorted other important actors such as GSEs. The medium of exchange for the system are reserve balances.
The Fed’s own work makes the remarkable claim that a change in domestic FedWire daily volumes display a 77% correlation coefficient with GDP, with a 4 quarter lag. As they point out such a high correlation coefficient is ‘suggestive of a relationship between GDP and large-value payments’ (via FedWire). The following chart and table are taken from that report.
The report uses data from 2004 till 2020 (Q1), so misses the massive increase in reserves and FedWire turnover that followed COVID19. We updated the work in the Fed paper to cover 2014:2025 (the extent of our comprehensive data). Our results show the correlation coefficient for FedWire Funds Service against NGDP remains high at 62%, though the lag has extended to 5 quarters (rather than 4).
Insight: there is long-running evidence that turnover of reserves (the liability to QE) via FedWire has an impact on nominal GDP – and almost certainly inflation.
That’s not all. Our earlier work showed the strong coincident relationship between FedWire turnover and average daily volume in securities and repo markets, with both closely linked to quantity of reserve balances. Indirectly, we can infer the turnover of FedWire, securities and repo is translated into wider economy transactions (GDP) with a 4-5 quarter lag.
The relationship between ADV of equities and FedWire is evident in the normalised (standard) representations below for equities (chart below). I’ve included the level of reserve balances to show the strong link to both series.
As reserve balances decline, their availability for settlement purposes becomes tighter, requiring higher daily velocity (turnover).
More analysis below.